A carregar...
Publicação
High-throughput automatized reach-to-grasp setup to study motor control and movement patterns
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 22.96 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Research on skilled motor control in rodents often faces significant limitations, including low trial throughput, heavy reliance on manual supervision, and limited ecological validity. This work addresses these challenges through the design and validation of a high-throughput, automated system for measuring skilled forelimb movements in mice. The platform was engineered to operate continuously in a homecage environment while preserving experimental rigour and animal welfare, thereby enabling long-term, minimally supervised behavioural studies. The reach-to-grasp task was chosen as the benchmark behaviour because it reliably evokes precise and repeatable forelimb movements, providing a stringent test of hardware and analytical performance. A custom reaching chamber was developed and progressively refined to automate water delivery, video capture, and data logging. Infra-red illumination and a single high-speed camera, combined with a 45° mirror, ensured stable, high-resolution recordings suitable for kinematic analysis. The resulting motiontracking pipeline, based on DeepLabCut, accurately extracted trajectories and derived metrics such as velocity, acceleration, and movement smoothness. To extend the task to a 24/7 setting, the chamber was integrated with the home-cage through a door mechanism featuring radio frequency identification (RFID) and a passive stopper system. This solution guarantees reliable individualisation and prevents multiple entries without introducing stress, highlighting the importance of aligning engineering design with natural behaviour. Pilot experiments confirmed the system’s stability and compatibility with parallel physiological recordings, including electromyography and in vivo calcium imaging. Although task-specific dystonia served as a proof-of-concept application, the platform itself is disease-agnostic. Its modular architecture, automated control, and scalable data acquisition provide a robust foundation for longitudinal studies of motor learning, adaptation, and pathology, bridging the gap between tightly controlled laboratory tasks and naturalistic behaviour
Descrição
Tese de Mestrado, Engenharia Biomédica e Biofísica, 2026, Universidade de Lisboa, Faculdade de Ciências
Palavras-chave
High-throughput behaviour Motion tracking Motor control Reach-to-grasp Task-specific dystonia